Merge branch 'main' into dbt-cortex-search-service-20260614

Author: Matts52

.github/pull_request_template.md

@@ -7,6 +7,12 @@ Link to your package's repository:
 ## Checklist
 _This checklist is a cut down version of the [best practices](../blob/main/package-best-practices.md) that we have identified as the package hub has grown. Although meeting these checklist items is not a prerequisite to being added to the Hub, we have found that packages which don't conform provide a worse user experience._
 
+### Real world usage
+
+Using coding agents to help you build your package is totally fine, but please don't submit it to the package hub until you've used it in your own production environment for an extended period of time.
+
+- [ ] (Required): I have been using this package in production and am satisfied with its behavior.
+
 ### First run experience
 - [ ] (Required): The package includes a [licence file detectable by GitHub](https://docs.github.com/en/communities/setting-up-your-project-for-healthy-contributions/adding-a-license-to-a-repository), such as the Apache 2.0 or MIT licence.
 - [ ] The package includes a README which explains how to get started with the package and customise its behaviour

.github/workflows/hubcap-scheduler.yml

@@ -66,6 +66,7 @@ jobs:
           HUBCAP_CONFIG: ${{ secrets.HUBCAP_CONFIG }}
           INPUT_DRY_RUN: ${{ github.event.inputs.dry_run || (github.event_name == 'pull_request' && 'true') || 'false' }}
           ENVIRONMENT: ${{ github.event.inputs.environment || (github.event_name == 'pull_request' && 'test') || 'production' }}
+          FUSION_BINARY: '/home/runner/.local/bin/dbt'
         run: |
           echo "Using $ENVIRONMENT environment configuration"
 
@@ -90,13 +91,16 @@ jobs:
             echo "$HUBCAP_CONFIG" > base_config.json
           fi
 
+          # Add Fusion binary path from env var
+          jq --arg fusion_path "$FUSION_BINARY" '.fusion_binary_path = $fusion_path' base_config.json > base_config_fusion.json
+
           # Modify config for dry run if requested
           if [ "$INPUT_DRY_RUN" = "true" ]; then
             echo "Enabling dry-run mode (push_branches = false)"
-            jq '.push_branches = false' base_config.json > config.json
+            jq '.push_branches = false' base_config_fusion.json > config.json
           else
             echo "Running in live mode (push_branches = true)"
-            cp base_config.json config.json
+            cp base_config_fusion.json config.json
           fi
 
           # Set the CONFIG environment variable for hubcap.py
@@ -115,9 +119,10 @@ jobs:
           echo "=== Execution Summary ==="
 
           # Check for errors
-          if grep -q 'ERROR' hubcap.log; then
+          # Space after ERROR differentiates error logs from parse errors
+          if grep -q 'ERROR ' hubcap.log; then
             echo "❌ Found ERROR events in execution:"
-            grep 'ERROR' hubcap.log
+            grep 'ERROR ' hubcap.log
             echo "EXECUTION_STATUS=failed" >> $GITHUB_ENV
           else
             echo "✅ No ERROR-level log entries found"

CONTRIBUTING.md

@@ -29,13 +29,15 @@ Save the token to a secure location. Click the "Configure SSO" button and "Autho
 ```shell
 cp config.example.json config.json
 
-# Add the relevant GitHub username, email address and token
+# Add the relevant GitHub username, email address, and token
 $EDITOR config.json
 
 # Export the JSON credentials into an environment variable
 export CONFIG=$(<config.json)
 ```
 
+If you have Fusion installed in a location other than the default (`~/.local/bin/dbt`), you will also need to add a line to `config.json` that sets `fusion_binary_path` to the correct path.
+
 ## Run in test mode
 
 ### Local testing

hub.json

@@ -72,6 +72,9 @@
     "henryupton": [
         "dbt-scd2-utils"
     ],
+    "kanoniv": [
+        "dbt-kanoniv"
+    ],
     "LewisDavies": [
         "upstream-prod"
     ],
@@ -91,7 +94,8 @@
         "dbt-set-similarity",
         "dbt-stat-test",
         "dbt-cortex-search-service",
-        "dbt-ml-eval"
+        "dbt-ml-eval",
+        "dbt-orphan"
     ],
     "metaplane" : [
         "dbt-expectations"
@@ -350,12 +354,17 @@
         "dbt_ga4_export",
         "dbt_unified_rag",
         "dbt_amazon_selling_partner",
-        "dbt_dynamics_365_crm"
+        "dbt_dynamics_365_crm",
+        "dbt_stackadapt"
     ],
     "fullstorydev": [
         "dbt_fullstory"
     ],
+    "gauthierpiarrette": [
+        "dbt-timefence"
+    ],
     "get-select": [
+        "dbt-query-tags",
         "dbt-snowflake-monitoring",
         "dbt-snowflake-query-tags"
     ],
@@ -426,6 +435,9 @@
     "omnata-labs": [
         "dbt-ml-preprocessing"
     ],
+    "paradime-io": [
+        "dbt-llm-evals"
+    ],
     "pvcy": [
         "dbt-privacy"
     ],
@@ -450,6 +462,9 @@
     "sgoley": [
         "dbt-postgres-utils"
     ],
+    "smithclay": [
+        "dbt_opentelemetry"
+    ],
     "snowplow": [
         "dbt-snowplow-utils",
         "dbt-snowplow-web",
@@ -482,12 +497,20 @@
         "dbt-meta-testing",
         "dbt-activity-schema"
     ],
+    "tripleaceme": [
+        "dbt-anomaly-detector",
+        "dbt-doc-inherit",
+        "dbt-doc-tracker"
+    ],
     "tuva-health": [
         "the_tuva_project"
     ],
-    "versent-dimensional": [
+    "versent": [
         "dbt_versent_dimensional_automate"
     ],
+    "vladimir-vvalov": [
+        "tbmacro"
+    ],
     "windsor-ai": [
         "dbt_facebook_ads",
         "dbt-bigquery-package-for-salesforce",
@@ -507,5 +530,8 @@
     ],
     "datooine": [
         "dbt-buckets"
+    ],
+    "teoria": [
+        "dbt-iceberg-utils"
     ]
 }

hubcap/exceptions.py

@@ -0,0 +1,40 @@
+class GitOperationError(Exception):
+    """Custom exception for git operation failures"""
+
+    pass
+
+
+class ConfigurationError(Exception):
+    """Custom exception for configuration errors"""
+
+    pass
+
+
+class FileOperationError(Exception):
+    """Custom exception for file operation errors"""
+
+    pass
+
+
+class PackageMaintainerError(Exception):
+    """Custom exception for package maintainer loading errors"""
+
+    pass
+
+
+class PackageError(Exception):
+    """Custom exception for package operation failures"""
+
+    pass
+
+
+class ReleaseCarrierError(Exception):
+    """Custom exception for release carrier operation failures"""
+
+    pass
+
+
+class VersionError(Exception):
+    """Custom exception for version operation errors"""
+
+    pass

hubcap/git_helper.py

@@ -10,28 +10,23 @@
 from git import Repo
 from git.exc import GitCommandError
 from pathlib import Path
+from .exceptions import GitOperationError
 
 
-class GitOperationError(Exception):
-    """Custom exception for git operation failures"""
-
-    pass
-
-
-def run_cmd(cmd, quiet=False):
+def run_cmd(cmd: str, quiet=False):
     """Dispatching commands to shell from inside Python needs an exit status wrapper"""
     try:
-        proc = subprocess.run(
+        proc: subprocess.CompletedProcess[bytes] = subprocess.run(
             args=cmd.split(), capture_output=True, timeout=300
         )  # 5 minute timeout
 
         if proc.returncode:
-            stderr_output = proc.stderr.decode("utf-8").rstrip()
+            stderr_output: str = proc.stderr.decode("utf-8").rstrip()
             logging.warning(f"Command failed: {cmd}")
             logging.warning(f"Error: {stderr_output}")
             raise GitOperationError(f"Command failed: {cmd}. Error: {stderr_output}")
 
-        output = proc.stdout.decode("utf-8").rstrip()
+        output: str = proc.stdout.decode("utf-8").rstrip()
 
         if not quiet:
             logging.info(output)

hubcap/helper.py

@@ -8,18 +8,7 @@
 from typing import Dict, Any, List
 
 from pathlib import Path
-
-
-class ConfigurationError(Exception):
-    """Custom exception for configuration errors"""
-
-    pass
-
-
-class FileOperationError(Exception):
-    """Custom exception for file operation errors"""
-
-    pass
+from .exceptions import ConfigurationError, FileOperationError
 
 
 NOW = int(datetime.datetime.now().timestamp())

hubcap/hubcap.py

@@ -1,4 +1,5 @@
 import logging
+from hubcap.types import PathLike
 import sys
 from pathlib import Path
 
@@ -15,6 +16,9 @@
 )
 from hubcap.records import IndividualPullRequests, ConsolididatedPullRequest
 
+# Default if no path is specified in config.json
+DEFAULT_FUSION_BINARY_PATH = Path.home() / ".local" / "bin" / "dbt"
+
 
 def app():
     exit_code = main()
@@ -57,6 +61,9 @@ def main():
         git_tmp = "target"
         TMP_DIR = Path(git_tmp).resolve()
         selected_package_maintainers = config.get("selected_package_maintainers")
+        fusion_binary_path: PathLike = config.get(
+            "fusion_binary_path", DEFAULT_FUSION_BINARY_PATH.resolve()
+        )
 
         try:
             PACKAGE_MAINTAINERS = package_maintainers.load_package_maintainers()
@@ -156,7 +163,11 @@ def main():
         logging.info("collecting the new version tags for packages checked into hub")
         try:
             update_tasks = package.get_update_tasks(
-                PACKAGE_MAINTAINERS, HUB_VERSION_INDEX, TMP_DIR, github_repo
+                PACKAGE_MAINTAINERS,
+                HUB_VERSION_INDEX,
+                TMP_DIR,
+                github_repo,
+                fusion_binary_path,
             )
         except Exception as e:
             logging.error(f"Error getting update tasks: {str(e)}")

hubcap/package.py

@@ -6,17 +6,13 @@
 from typing import List, Dict, Any
 
 from git import Repo
+from hubcap.types import PathLike
 from pathlib import Path
 
 from hubcap import records
 from hubcap import version
 from hubcap.git_helper import clone_repo, run_cmd, GitOperationError
-
-
-class PackageError(Exception):
-    """Custom exception for package operation failures"""
-
-    pass
+from hubcap.exceptions import PackageError
 
 
 def clone_package_repos(package_maintainer_index, path):
@@ -50,7 +46,7 @@ def clone_package_repos(package_maintainer_index, path):
     return failed_repos
 
 
-def parse_pkg_name(repo_dir) -> str:
+def parse_pkg_name(repo_dir: Path) -> str:
     """Parse package name from dbt_project.yml with error handling"""
     try:
         dbt_project_path = repo_dir / Path("dbt_project.yml")
@@ -160,7 +156,13 @@ def parse_pkgs(repo_dir) -> List[Dict[str, Any]]:
         return []
 
 
-def get_update_tasks(maintainers, version_index, path, hub_repo):
+def get_update_tasks(
+    maintainers: list[records.PackageMaintainer],
+    version_index,
+    path: Path,
+    hub_repo,
+    fusion_binary_path: PathLike,
+):
     """build list of tasks for package version-bump commits"""
 
     def has_dbt_project_yml(package, directory):
@@ -170,10 +172,10 @@ def has_dbt_project_yml(package, directory):
             logging.warning(f"{package} has no dbt_project.yml. Skipping...")
         return has_yaml
 
-    def get_new_tags(repo_path, maintainer_name):
+    def get_new_tags(repo_path: Path, maintainer_name: str):
         # Existing tags are fetched from version index
         try:
-            yml_package_name = parse_pkg_name(repo_path)
+            yml_package_name: str = parse_pkg_name(repo_path)
             logging.info(f"collecting tags for {yml_package_name}")
 
             existing_tags = version.get_existing_tags(
@@ -195,7 +197,7 @@ def get_new_tags(repo_path, maintainer_name):
 
     def build_update_task_tuple(maintainer_name, hub_package_name):
         SKIP = None
-        repo_path = path / Path(f"{maintainer_name}_{hub_package_name}")
+        repo_path: Path = path / Path(f"{maintainer_name}_{hub_package_name}")
 
         # Check if repo directory exists
         if not repo_path.exists():
@@ -226,6 +228,7 @@ def build_update_task_tuple(maintainer_name, hub_package_name):
                     existing_tags=existing_tags,
                     new_tags=new_tags,
                     hub_repo=hub_repo,
+                    fusion_binary_path=fusion_binary_path,
                 )
             # Cannot create update task for package without new tags
             else:
@@ -266,7 +269,7 @@ def commit_version_updates_to_hub(
 
             # good house keeping
             os.chdir(hub_dir_path)
-            cmd = f"git checkout {default_branch}"
+            cmd = f"git checkout --force {default_branch}"
             run_cmd(cmd)
         except Exception as e:
             logging.error(

hubcap/package_maintainers.py

@@ -4,12 +4,7 @@
 from pathlib import Path
 
 from hubcap.records import PackageMaintainer
-
-
-class PackageMaintainerError(Exception):
-    """Custom exception for package maintainer loading errors"""
-
-    pass
+from hubcap.exceptions import PackageMaintainerError
 
 
 def load_package_maintainers() -> List[PackageMaintainer]: